The Interactive Network Visualization of the Interactions Between

Topologically Associating Domains in the Genome of Fruit Fly

Samira Mali

1 a

and Swetha Annavarapu

2,3 b

Laboratory for Theoretical and Computational Molecular Biophysics, Virginia Tech, Blacksburg VA, U.S.A.

The Distributed Virtual Environments Lab, Center for Human-Computer Interaction, Virginia Tech, Blacksburg VA, U.S.A.

ADP Inc., Alpharetta GA, U.S.A.

Keywords:

Topologically Associating Domain, Genome of Fruit Fly, Human-Centered Computing, Visualization

Technique, Network Visualization, Visualization Design, Evaluation Methods.

Abstract:

In this work, we created a network visualization to help you understand how Topologically Associating Do-

mains (TADs) interact with each other across the genome based on where the TAD is located, whether it is near

the center of the nucleus or near the edge of the nucleus. This visualization can reveal how the dense regions

and sparse regions of chromosome interactions are distributed in one view. The pilot study demonstrates how

network visualization of TAD-TAD interactions can quickly answer numerous major questions in 3D genome

and epigenetics ﬁeld without requiring the development of Machine Learning methods or Algorithms to un-

lock Heatmap structures. The questions include but are not limited to, determining many-way interactions and

interactions between TADs belonging to various epigenetic classes.

1 INTRODUCTION

1.1 Topologically Associating Domain

Several researchers are studying how DNA folds in

the tiny nucleus space. Building artiﬁcial chromo-

somes will successfully correct genetic abnormalities

if they follow the correct folding other than the correct

orders across the genome. Researchers now know that

folding is not random, and there are some patterns we

can observe in the folding of chromosomes wrapped

around the Histone proteins. Topologically Associ-

ating Domains (TAD)(Lieberman-Aiden et al., 2009)

are one of those basic folding patterns.

A topologically associating domain (TAD) is a

self-interacting genomic region in which DNA se-

quences interact with one another more frequently

than sequences outside the TAD.

TADs are structural features of chromosomes that

play a crucial role in genome organization and regu-

lation of gene expression. They were ﬁrst identiﬁed

in 2012 in Drosophila melanogaster and subsequently

in mammalian genomes using high-throughput chro-

mosome conformation capture (Hi-C) technology.

https://orcid.org/0000-0002-3701-0072

https://orcid.org/0000-0002-1221-2823

(Dixon et al., 2012)).

TADs have been found to be conserved across

species and are thought to be important for main-

taining proper gene expression and regulatory inter-

actions(Acemel et al., 2017). TADs are thought to be

stable over time, and disruption of TAD boundaries

has been associated with various diseases, including

cancer and developmental disorders (Flavahan et al.,

2016)). TAD structure has an essential role in gene

regulation because the TAD boundaries show the ex-

act position of insulator proteins. Disruption of TAD

boundaries is found to be associated with a wide range

of diseases, such as cancer.(Bonev and Cavalli, 2016)

As a result, recognizing TADs and any information

about them would be very helpful in the 3D genome

ﬁeld.

TADs are believed to be formed by the binding of

architectural protein complexes, such as CTCF and

cohesin, which loop the DNA to form discrete do-

mains (Nora et al., 2012). Recent studies have shown

that TADs are not always strictly compartmentalized

but can interact with each other, leading to the forma-

tion of so-called ”meta-TADs” (Bonev and Cavalli,

2016). In addition, several complexes, such as CTCF

and the cohesin protein complex, are recognized for

their connection to the creation of TADs. These two

main protein complexes have a role in the folding of

504

Mali, S. and Annavarapu, S.

The Interactive Network Visualization of the Interactions Between Topologically Associating Domains in the Genome of Fruit Fly.

DOI: 10.5220/0012456200003657

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2024) - Volume 1, pages 504-511

ISBN: 978-989-758-688-0; ISSN: 2184-4305

chromosomes. All this information conﬁrms that the

TADs are important and need to be explored in detail.

There is ongoing research to better understand the dy-

namics of TAD formation and maintenance, as well as

their role in gene regulation and disease (Dekker and

Mirny, (Dekker and Mirny, 2016))

1.2 Importance of Network

Visualization

Visualizing TAD (Topologically Associating Do-

mains) interactions is an important aspect of 3D

genome visualization, and heatmaps are one of the

most commonly used techniques for this purpose.

Heatmaps are used to represent the frequency of in-

teractions between different TADs, with each row

and column corresponding to a different TAD and the

color indicating the frequency of the interaction in the

population of cells.

However, as with any visualization technique,

there are limitations to using heatmaps for TAD inter-

action visualization(Lex et al., 2012). Not only may

heatmaps not provide a clear sense of the spatial orga-

nization of TADs in three-dimensional space, but also

one of their limitations is that they can only show pair-

wise interactions between TADs, so they may not be

able to capture more complex relationships between

multiple TADs.

There are no 3D visualizations for Many-Body

chromatin interactions and with the regular heatmaps,

it is difﬁcult to ﬁnd the three-way or more-way in-

teractions(Oudelaar et al., 2018). Regarding this

concept, to the best of our knowledge, there is

only one work available for Drosophila(Sun et al.,

2021), whose resolution is higher than TAD. On the

other hand, there is a recent research (Dotson et al.,

2022) about the many-body interactions in the human

genome that used upset plots to show the intersections

of the pairwise interactions to recognize many-body

interactions.

We argue that network visualization would be

more effective than an upset plot in showing the inter-

section among data. Upset plots are generally more

effective for comparing the overlap of sets or groups

of data. However, network visualizations are more

useful for understanding the relationship between en-

tities. In this work, we propose such a network visu-

alization of TADs.

One of our speciﬁc ideas in this research is ad-

dressing the issue of the incompleteness of many-

body contact view in heatmap by using network

visualization of interactions between TADs. In ad-

dition, Hi-C heatmaps do not directly represent the

directionality of interactions between genomic seg-

ments (TADs), which can make it difﬁcult to iden-

tify the nature of the interactions. Network visualiza-

tion can address this by incorporating directed graphs

(edges).

Other visualization techniques, such as 3D plots

and network graphs, can be used in conjunction with

heatmaps. 3D plots can be used to show the physi-

cal arrangement of TADs in three-dimensional space,

while network graphs can be used to show the rela-

tionships between multiple TADs in a more complex

and nuanced way.

Overall, while heatmaps can be a useful tool for

visualizing TAD interactions, they may need to be

supplemented with other visualization techniques to

capture the complexity of 3D genome organization

fully.

On the other hand, circus plots can be used to visu-

alize the TAD-TAD interactions. However, it has sev-

eral disadvantages, such as limited scalability, mean-

ing these plots can become cluttered and difﬁcult to

read when the number of individuals or interactions

is large as well in the circus plots, it can be difﬁcult to

identify the speciﬁc connections or relationships be-

tween individuals, particularly when there are many

overlapping lines. All of these issues can be addressed

easily by a well-rendered graph layout visualization.

Furthermore, one important aspect of the

structure-function relationship in chromatin in the

nucleus is the interaction between active and inac-

tive regions of chromatin. Therefore, our second

speciﬁc idea in this project is to propose a network

visualization to represent types of interactions,

including interactions among Active, PcG, HP1

Centromeric, and Null TADs in the genome of fruit

ﬂy.

Network visualization, also known as graph vi-

sualization, is a technique used to display complex

data in a way that is easy to understand. Networks

are made up of nodes and edges, with nodes rep-

resenting objects or entities and edges representing

relationships between them. In the context of TAD

visualization, nodes can represent individual TADs,

while edges can represent interactions between TADs.

There are several tools available for visualizing TADs,

which can be broadly classiﬁed into two categories:

1) tools for visualizing TADs as 2D plots and 2)

tools for visualizing TADs in 3D. Tools for visualiz-

ing TADs as 2D plots:

• Juicebox (Durand et al., 2016) is a popular tool for

visualizing Hi-C data, which can be used to dis-

play TAD boundaries, as well as other genomic

features such as gene locations and epigenetic

marks.

• HiCPlotter (Akdemir and Chin, 2015) is a tool

The Interactive Network Visualization of the Interactions Between Topologically Associating Domains in the Genome of Fruit Fly

505

for visualizing Hi-C data as 2D heatmaps, which

can be used to identify TAD boundaries and other

structural features of the genome.

Tools for visualizing TADs in 3D:

• TADview 1.1 is a tool for visualizing TADs as 3D

models. This tool is a plugin that can be added to

VMD (Humphrey et al., 1996) and includes many

features, such as representing different epigenetic

classes of TADs.

• 3D Genome Browser is a web-based tool for vi-

sualizing Hi-C data in 3D, which can be used to

explore TAD boundaries and their spatial organi-

zation. (Wang et al., 2018)

• HiCExplorer (Wolff et al., 2018): A suite of tools

for exploring Hi-C data, including a 3D visualiza-

tion tool that can be used to visualize TADs and

other genomic features in 3D.

• Chrom3D(Paulsen et al., 2017) is a computational

tool for modeling the 3D structure of the genome,

which can be used to visualize TADs and other

structural features in 3D.

In High-throughput chromatin conformation cap-

ture (Hi-C) experiments, these regions are basically

visualized as squares along the diagonals through a

Heat Map in higher resolution. While in TAD reso-

lution, each dot in the heatmap shows one TAD. Heat

Maps have a lot of limitations in terms of visualization

context. They only show the frequency of interactions

in each region and cannot show the types of the TAD.

Other than epigenetic classes, TADs can be classiﬁed

into two categories based on their position with re-

spect to the nucleus, whether they are near the nuclear

envelope or near the center of the nucleus(Afanasyev

and Onufriev., 2022). There are several ways in which

network visualization can be used to enhance TAD vi-

sualization. For example, it can be used to:

1. Display relationships between TADs: By using a

network visualization, it is possible to see the re-

lationships between TADs and how they interact

with each other. This can provide a more compre-

hensive view of the genome than a static image of

TADs alone.

2. Analyze the structure of TADs: Network visual-

ization can be used to analyze the structure of

TADs and identify patterns and clusters within

them. This can help to uncover functional rela-

tionships and potential regulatory mechanisms.

3. Identify key TADs: By using network analysis

techniques, it is possible to identify key TADs that

are important for the overall structure and func-

tion of the genome.

2 METHODOLOGY USED TO

COLLECT PAPERS

The keywords used to collect the papers were

”Topologically Associating Domains,” ”TAD visu-

alization,” ”TAD Network Visualization,” ”Genome

Network Visualization,” and ”Network Visualization

Tools for Biological Networks.” A total of 48 papers

were reviewed out of which 29 papers were chosen as

references for the literature review.

The reason for choosing these papers is that they

are aligned with two main goals of this study, in-

cluding the interaction among the different epigenetic

classes of chromosomes as well as the many-body in-

teraction concept in 3D genome folding.

Three papers were published in Genome Biol-

ogy journal, four papers in Nucleic acids research,

six papers in Nature (communications, review, genet-

ics), two papers in Cell Journal, one from Epigenet-

ics and Chromatin journal, one from Interdisciplinary

Reviews: Developmental Biology, one in Bioinfor-

matics journal, one was accepted in Proceedings of

the international AAAI conference on web and so-

cial media, one in Communications Biology journal,

one in scientiﬁc report, one journal in PLoS Compu-

tational Biology journal and one from Genome Re-

search journal.

3 RESEARCH GAPS

A few research gaps identiﬁed based on the current

visualization methods being used for TADs are:

1. There are not any 3D visualizations for Many-

Body chromatin interactions. With the regular

heatmaps, it is difﬁcult to show the three-way or

more-way interactions.(Liu et al., 2021)(Dotson

et al., 2022)

2. So far, heat maps are being used for TAD visual-

ization, but it does not give full information about

the interactions. There are only values of interac-

tions. But with a Network visualization we can

see if there is any interaction between the nodes

based on edge connection and its color taxonomy.

This project aims to ﬁll a gap in research by visu-

alizing interactions between TADs (topologically as-

sociated domains) in the nucleus. By mapping how

these interactions vary based on TAD location (near

the nucleus periphery or center), it can shed light on

gene activity patterns crucial for understanding gene

regulation. This visualization could aid physicist.

BIOINFORMATICS 2024 - 15th International Conference on Bioinformatics Models, Methods and Algorithms

506

4 HYPOTHESIS

We argue that a heatmap alone cannot depict the

distribution of TAD-TAD interactions in the nucleus

based on the epigenetic classes of TADs. Further-

more, the heatmap fails to represent multiple-body

interactions, making it challenging to discern 3-4-5

body interactions quickly and easily.

Based on both the design procedures we devel-

oped and the pilot study we conducted, we demon-

strate heatmaps are not enough to answer the afore-

mentioned question rapidly.

5 METHODS

5.1 Network Visualization of TAD-TAD

Interactions

In the proposed approach, each node represents a

TAD in the network, and each edge shows the inter-

action between the TADs. If there is an interaction

between the TADs, an edge is connected to repre-

sent it. Otherwise, there is no edge between them.

The TADs close to the nuclear periphery are inac-

tive regions of chromatin and those close to the center

are active regions of chromatin. There are 4 types

of nodes: Active, PcG (repressed Polycomb group),

HP1 Centromeric (inactive regions related to H3K4

histone modiﬁcation), and Null TADs. The size of

the node represents the radius of the TAD. TADs con-

tain information about the amount of DNA base-pair

on that domain as well as the exact region for that do-

main along the genome.

The initial plan was to use some available tools,

such as cola.js(Dwyer, 2018), to build an interactive

network visualization. However, due to some techni-

cal issues relating to a high number of nodes, Gephi

(Gansner and North, 2000) was used for the network

visualization. The other options, including cola.js and

plotly (Python library), do not meet our needs in terms

of applying the appropriate types of forces between

nodes to handle very large networks in biology con-

text. This resulted in these tools not being able to

accommodate the high number of nodes used to rep-

resent TADs across the whole genome.

5.2 Dataset

Generally, the data corresponding to the TAD loca-

tions is deﬁned by applying some predeﬁned algo-

rithms to Hi-C data. However, the data for this project

is taken from (Sexton et al., 2012), which partitioned

Figure 1: Many-Body Interaction in TAD Network Visual-

ization.

Figure 2: Zoomed-in view of a part of the Network. We can

zoom in and zoom out through the network.

the genome of Drosophila into Topologically Associ-

ating Domains (TAD) for each chromosome based on

the chromatin density in the nucleus.

This data is imported into the software interface

as a CSV ﬁle. We use Gephi 1.10(Bastian et al.,

2009) to visualize our TAD-TAD network. We use

this tool because it uses forces and other network fea-

tures to make it more readable. Several studies used

Gephi for their visualization targets. (Deng et al.,

2022),(Ma et al., 2021). We have used two parts of

the dataset: the ﬁrst is for the whole chromosomes of

the drosophila genome, and the second is only for Chr

X:[ 173,850-21,889,749] that has 184 TADs. Each

The Interactive Network Visualization of the Interactions Between Topologically Associating Domains in the Genome of Fruit Fly

507

Table 1: Each element represents the frequency of contacts

between each of the two domains.

1 2 3 4 5 6 7 8 9 10

1 1 0.999667 0.749417 0.555148 0.323559 0.216928 0.223259 0.183272 0.080973 0.0816395

2 0.999667 1 0.999334 0.874375 0.518161 0.357547 0.339553 0.257581 0.108964 0.109963

3 0.749417 0.999334 1 0.999334 0.691103 0.479174 0.423192 0.296235 0.112962 0.115961

4 0.555148 0.874375 0.999334 1 0.999334 0.870377 0.718427 0.487171 0.194269 0.186604

5 0.323559 0.518161 0.691103 0.999334 1 1 0.977341 0.647784 0.250916 0.224592

6 0.216928 0.357547 0.479174 0.870377 1 1 1 0.781073 0.272909 0.243585

7 0.223259 0.339553 0.423192 0.718427 0.977341 1 1 0.991003 0.460846 0.401866

8 0.183272 0.257581 0.296235 0.487171 0.647784 0.781073 0.991003 1 0.999334 0.952349

9 0.080973 0.108964 0.112962 0.194269 0.250916 0.272909 0.460846 0.999334 1 1

10 0.0816395 0.109963 0.115961 0.186604 0.224592 0.243585 0.401866 0.952349 1 1

Table 2: The attributes for each TAD include radius, epi-

class, as well as start and end loci.

EpiClass Start-EndLoc Radius (nm)

1 Inactive [173850 - 425249] 101.24

2 PcG [425250 - 513049] 71.29

3 Active [513050 - 551249] 54.02

4 Inactive [551250 - 658949] 76.32

5 Inactive [658950 - 1129149] 124.73

6 Active [1129150 - 1268749] 83.21

7 Active [1268750 - 1293949] 47.03

8 Inactive [1293950 - 1356849] 63.79

9 Active [1356850 - 1375149] 42.27

10 Active [1375150 - 1410649] 52.72

11 Inactive [1410650 - 1558649] 84.85

12 Inactive [1558650 - 1753549] 93

13 Active [1753550 - 1971649] 96.56

14 Active [1971650 - 2013749] 55.8

15 Active [2013750 - 2117549] 75.39

16 Active [2117550 - 2165549] 58.3

matrix element shows the edge weight (the frequency

of interaction among the population of cells).

5.3 Design Procedure

We imported a CSV ﬁle as a matrix with timestamps,

choosing graph settings like undirected edges, auto-

scaling, and excluding self-loops. Additionally, we

selected a sum-based edge merge strategy. Later, we

added node data (TAD labels, epigenetic classes, lo-

cations, and radii) to the workspace, merging dupli-

cates based on attributes like epiclass and location.

In Gephi’s Data Laboratory, we appended this data

to an existing workspace. We segregated epigenetic

classes using ﬁlters and identiﬁed Many-Body inter-

actions between active nodes. We also set up visual

aids, assigning node colors by epiclass and adjusting

node sizes based on radius. Edge colors were unique

for connecting TADs.

To inspect nodes easily, we installed the Inspec-

tor plugin, enabling us to view node details (ID, label,

timestamp, epiclass, locations, and radius) when hov-

ering over nodes in the graph area of the Overview

window.

In the overview window under the Layout tab, we

ran ForceAtlas 2(Jacomy et al., 2014) with the follow-

Figure 3: By hovering over a node, we can see the Node

information.

Figure 4: The chromosome shown by the circle is the chro-

mosome X amongst the whole chromosomes of Drosophila.

In addition, in this ﬁgure, we can observe some graph fea-

tures, such as clique, which could be directly related to the

higher level of gene expression in those chromosomal re-

gions.

ing speciﬁcations:

We have created two different ﬁlters: the ﬁrst is to

show all the Epigenetic classes, and the second is to

ﬁlter out connections between active nodes to identify

Many-Body interactions.

For Epigenetic Classes ﬁlter, we ﬁrst added the

intersection operator in the queries. The edge weight

ﬁlter ranged from the least edge weight above zero to

2. We adjusted the edge weight ﬁlter at 0.8 (Sun et al.,

2021).

For Many-Body Interactions ﬁlter, we created a

base ﬁlter similar to the epigenetic classes ﬁlter. We

modiﬁed the Inter edge of the Epiclass ﬁlter in queries

by choosing Active and PcG classes.

BIOINFORMATICS 2024 - 15th International Conference on Bioinformatics Models, Methods and Algorithms

508

Figure 5: (a)ForceAtlas 2 is developed by the Gephi team

as an all-around solution to Gephi users’ typical networks

(scale-free, 10 to 10,000 nodes), (b) The empirical studies

show the 0.8 is the distance threshed for having contact be-

tween pieces of DNA, (c) Many-Body Interaction ﬁlter in

TAD Network Visualization with Inter edges between Ac-

tive and PcG Nodes

Whenever we want to apply a ﬁlter, we can choose

that particular ﬁlter from the queries and click on the

Filter button at the bottom right corner to update it in

the network.

The reason we apply the ForceAtlas2 is: we need

to apply the right force to reach a smooth network

of TAD-TAD interaction in the Drosophila genome.

Previous studies provide different models. One pro-

vides statistical models on Force-directed graph lay-

out(Zhong et al., 2023), and the other uses the uniﬁed

force representation integrating popular techniques

such as the stress model, the spring-electrical model,

as well as maxent-stress model(Xue et al., 2022).

However, ForceAtlas2 adjusts attractive forces in pro-

portion to distance and repulsive forces inversely to

distance, aiming to enhance the display of local neigh-

borhoods and cluster structures in the visualization,

which is more similar to the (Tolokh and Onufriev,

2023) which is a model simulation for Drosophila

(fruit ﬂy) genome.

5.4 Initial Pilot Study

To prepare a better pilot study, we used the ideas of

a few pilot study works (Yang and Goodwin, 2019),

(Sedlmair et al., 2012), (Wall et al., 2022). We present

our analysis of two expert interviews. Such analysis

is crucial to understanding the real-world scenarios of

analyzing chromatin organization in TAD resolution.

The motivation for these interviews was to under-

stand the role of ﬂow data in real-world applications

and existing workﬂows across different disciplines to

help inform new visualization designs.

The questions for the pilot study are listed in the

link:

Pilot Study Questions.

A few guidelines formulated based on the pilot

study are:

1. In the heatmap, the checkerboard patterns show

the active and inactive regions. However, deciding

which one is active and which one is inactive is

sometimes confusing.

The distribution of four types of Epigenetic

classes of TADs can be clear in the network, with

different color coding for each class.

2. One of the major challenges in single-cell stud-

ies is ﬁguring out what types of single cells are

more frequent in the ensemble average Hi-C con-

tact matrices. We cannot answer this question ex-

plicitly using heatmap visualization.

However, we can do this using the information

about the degree of each node in the network.

3. Heatmap can be used to identify the many-body

interactions, but it takes a lot of time.

Network visualization can make the process of

identifying many-body interactions easier.

4. The most important one is that some graph fea-

tures like clique show a high level of transcription

and gene expression in those regions (hubs), and

we can get these features by network visualization

(see Figure 5).

The results of our analysis show that network vi-

sualization can be more efﬁcient and easier to use than

interaction matrices (Hi-C heatmaps), which are the

most common approaches to visualizing chromatin

interactions in this ﬁeld.

The Interactive Network Visualization of the Interactions Between Topologically Associating Domains in the Genome of Fruit Fly

509

6 RESULTS AND FUTURE PLAN

FOR USER STUDY

We develop a pipeline for network visualization for

TAD-TAD interaction of Drosophila genome using a

graph layout method providing ﬁlters such as multi-

way interactions as well as epigenetic classes. The re-

sults from the pilot study highlight the need of a new

visualization approach for TAD-TAD interactions to

facilitate answering several questions: ”Visualizing

networks simpliﬁes the identiﬁcation of many-body

interactions”. In addition, the degree of the nodes

in the network can explain the biological meaning of

this type of representation of interactions between the

TADs: ”The integration of the nodes with high de-

grees show the hub regions with higher level of tran-

scription or gene expression.”

The combination of our design procedure and pi-

lot study demonstrate the correctness of our hypothe-

sis which indicate Hi-C heatmap in TAD resolution is

not enough to smoothly justify how TADs distributed

in different radial positions as well as quick determi-

nation of many-way interactions among TADs.

In the current work, primary focus for the project

is designing a Network Visualization for TADs. How-

ever, since this work is domain-speciﬁc, we plan on

conducting a user study in the future. As part of the

study, we plan to recruit 10 Ph.D. students in the Bi-

ology/Physics/Biochemistry ﬁelds, where we plan to

show them the visualizations and conﬁrm if these vi-

sualizations will be helpful in the ﬁeld using a survey.

As of now, the User studies are yet to be done

fully. As part of the user study, the participant would

be interacting with the network visualization. They

would be asked to perform general interaction tasks

such as zoom-in and zoom-out, identify active and in-

active TADs, and identify TAD many-body interac-

tions such as 3, 4, 5-body interactions. They will be

asked to list out the nodes that are part of the Many-

body interactions. In the end, they would be asked

to ﬁll out an online survey, which includes a NASA-

TLX questionnaire to understand the mental work-

load of using the network visualization and SUS to

understand the ease of system usability. The survey

will also include a few custom questions about the

TAD network visualizations to understand if the par-

ticipant could successfully perform all the tasks as-

signed to them during the study.

The results of the study will be analyzed based on

the answers provided by the participants to the survey

questions.

7 CONCLUSION

Brieﬂy, we can categorize our idea into two speciﬁc

aspects. First, we are interested in seeing how TADs

have interacted with each other in many ways. On

the other hand, we would like to see the distribution

of these interactions and compare them among the

Epigenetic classes Active, PcG (repressed Polycomb

group), HP1 Centromeric (inactive regions related to

H3K4 histone modiﬁcation), and Null in TADs. As

the future work, expanding the work to apply the tech-

nique to mammalian genomes would also be inter-

esting. Answering these questions is helpful to Bi-

ologists and Physicists for an easier understanding of

genome folding.

ACKNOWLEDGMENT

The authors extend their appreciation to Yalong Yang

and Alexey Onufriev for their guidance and support.

REFERENCES

Acemel, R. D., Maeso, I., and G

omez-Skarmeta, J. L.

(2017). Topologically associated domains: a success-

ful scaffold for the evolution of gene regulation in an-

imals. Wiley Interdisciplinary Reviews: Developmen-

tal Biology, 6(3):e265.

Afanasyev, Alexander Y., Y. K. I. S. T. I. V. S. and Onufriev.,

A. V. (2022). The probability of chromatin to be at the

nuclear lamina has no systematic effect on its tran-

scription level in fruit ﬂy. bioRxiv.

Akdemir, K. C. and Chin, L. (2015). Hicplotter integrates

genomic data with interaction matrices. Genome biol-

ogy, 16:1–8.

Bastian, M., Heymann, S., and Jacomy, M. (2009). Gephi:

an open source software for exploring and manipu-

lating networks. In Proceedings of the international

AAAI conference on web and social media, volume 3,

pages 361–362.

Bonev, B. and Cavalli, G. (2016). Organization and func-

tion of the 3d genome. Nature Reviews Genetics,

17(11):661–678.

Dekker, J. and Mirny, L. (2016). The 3d genome as

moderator of chromosomal communication. Cell,

164(6):1110–1121.

Deng, L., Gao, B., Zhao, L., Zhang, Y., Zhang, Q., Guo, M.,

Yang, Y., Wang, S., Xie, L., Lou, H., et al. (2022). Di-

urnal rnapii-tethered chromatin interactions are asso-

ciated with rhythmic gene expression in rice. Genome

Biology, 23(1):1–22.

Dixon, J. R., Selvaraj, S., Yue, F., Kim, A., Li, Y., Shen, Y.,

Hu, M., Liu, J. S., and Ren, B. (2012). Topological

domains in mammalian genomes identiﬁed by analy-

BIOINFORMATICS 2024 - 15th International Conference on Bioinformatics Models, Methods and Algorithms

510

sis of chromatin interactions. Nature, 485(7398):376–

380.

Dotson, G. A., Chen, C., Lindsly, S., Cicalo, A., Dilworth,

S., Ryan, C., Jeyarajan, S., Meixner, W., Stansbury,

C., Pickard, J., et al. (2022). Deciphering multi-way

interactions in the human genome. Nature Communi-

cations, 13(1):5498.

Durand, N. C., Robinson, J. T., Shamim, M. S., Machol, I.,

Mesirov, J. P., Lander, E. S., and Aiden, E. L. (2016).

Juicebox provides a visualization system for hi-c con-

tact maps with unlimited zoom. Cell systems, 3(1):99–

101.

Dwyer, T. (2018). Constraint-Based Layout in the Browser.

https://marvl.infotech.monash.edu/webcola/.

Flavahan, W. A., Drier, Y., Liau, B. B., Gillespie, S. M.,

Venteicher, A. S., Stemmer-Rachamimov, A. O.,

Suv

a, M. L., and Bernstein, B. E. (2016). Insula-

tor dysfunction and oncogene activation in idh mutant

gliomas. Nature, 529(7584):110–114.

Gansner, E. R. and North, S. C. (2000). An open graph

visualization system and its applications to software

engineering. Software: practice and experience,

30(11):1203–1233.

Humphrey, W., Dalke, A., and Schulten, K. (1996). Vmd:

visual molecular dynamics. Journal of molecular

graphics, 14(1):33–38.

Jacomy, M., Venturini, T., Heymann, S., and Bastian, M.

(2014). Forceatlas2, a continuous graph layout algo-

rithm for handy network visualization designed for the

gephi software. PloS one, 9(6):e98679.

Lex, A., Streit, M., Schulz, H.-J., Partl, C., Schmalstieg,

D., Park, P. J., and Gehlenborg, N. (2012). Stratomex:

visual analysis of large-scale heterogeneous genomics

data for cancer subtype characterization. In Computer

graphics forum, volume 31, pages 1175–1184. Wiley

Online Library.

Lieberman-Aiden, E., Van Berkum, N. L., Williams, L.,

Imakaev, M., Ragoczy, T., Telling, A., Amit, I., La-

joie, B. R., Sabo, P. J., Dorschner, M. O., et al. (2009).

Comprehensive mapping of long-range interactions

reveals folding principles of the human genome. sci-

ence, 326(5950):289–293.

Liu, L., Zhang, B., and Hyeon, C. (2021). Extracting multi-

way chromatin contacts from hi-c data. PLoS Compu-

tational Biology, 17(12):e1009669.

Ma, J., Hou, C., Li, Y., Chen, S., and Wu, C. (2021).

Ogt protein interaction network (ogt-pin): a curated

database of experimentally identiﬁed interaction pro-

teins of ogt. International journal of molecular sci-

ences, 22(17):9620.

Nora, E. P., Lajoie, B. R., Schulz, E. G., Giorgetti, L.,

Okamoto, I., Servant, N., Piolot, T., Van Berkum,

N. L., Meisig, J., Sedat, J., et al. (2012). Spa-

tial partitioning of the regulatory landscape of the x-

inactivation centre. Nature, 485(7398):381–385.

Oudelaar, A. M., Davies, J. O., Hanssen, L. L., Tele-

nius, J. M., Schwessinger, R., Liu, Y., Brown, J. M.,

Downes, D. J., Chiariello, A. M., Bianco, S., et al.

(2018). Single-allele chromatin interactions identify

regulatory hubs in dynamic compartmentalized do-

mains. Nature genetics, 50(12):1744–1751.

Paulsen, J., Sekelja, M., Oldenburg, A. R., Barateau, A.,

Briand, N., Delbarre, E., Shah, A., Sørensen, A. L.,

Vigouroux, C., Buendia, B., et al. (2017). Chrom3d:

three-dimensional genome modeling from hi-c and

nuclear lamin-genome contacts. Genome biology,

18:1–15.

Sedlmair, M., Meyer, M., and Munzner, T. (2012). Design

study methodology: Reﬂections from the trenches and

the stacks. IEEE transactions on visualization and

computer graphics, 18(12):2431–2440.

Sexton, T., Yaffe, E., Kenigsberg, E., Bantignies, F.,

Leblanc, B., Hoichman, M., Parrinello, H., Tanay,

A., and Cavalli, G. (2012). Three-dimensional fold-

ing and functional organization principles of the

drosophila genome. Cell, 148(3):458–472.

Sun, Q., Perez-Rathke, A., Czajkowsky, D. M., Shao, Z.,

and Liang, J. (2021). High-resolution single-cell 3d-

models of chromatin ensembles during drosophila em-

bryogenesis. Nature communications, 12(1):205.

Tolokh, Igor S., N. A. K. I. V. S. and Onufriev, A. V. (2023).

Strong interactions between highly dynamic lamina-

associated domains and the nuclear envelope stabi-

lize the 3d architecture of drosophila interphase chro-

matin. Epigenetics & Chromatin, 16(1):21.

Wall, E., Xiong, C., and Kim, Y.-S. (2022). Vishikers’ guide

to evaluation: Competing considerations in study de-

sign. IEEE Computer Graphics and Applications,

42(3):29–38.

Wang, Y., Song, F., Zhang, B., Zhang, L., Xu, J., Kuang, D.,

Li, D., Choudhary, M. N., Li, Y., Hu, M., et al. (2018).

The 3d genome browser: a web-based browser for

visualizing 3d genome organization and long-range

chromatin interactions. Genome biology, 19(1):1–12.

Wolff, J., Bhardwaj, V., Nothjunge, S., Richard, G., Ren-

schler, G., Gilsbach, R., Manke, T., Backofen, R.,

Ram

ırez, F., and Gr

uning, B. A. (2018). Galaxy hic-

explorer: a web server for reproducible hi-c data anal-

ysis, quality control and visualization. Nucleic acids

research, 46(W1):W11–W16.

Xue, M., Wang, Z., Zhong, F., Wang, Y., Xu, M., Deussen,

O., and Wang, Y. (2022). Taurus: towards a uniﬁed

force representation and universal solver for graph

layout. IEEE Transactions on Visualization and Com-

puter Graphics, 29(1):886–895.

Yang, Y. and Goodwin, S. (2019). What-why analysis of ex-

pert interviews: Analysing geographically-embedded

ﬂow data. In 2019 IEEE Paciﬁc Visualization Sympo-

sium (PaciﬁcVis), pages 122–126. IEEE.

Zhong, F., Xue, M., Zhang, J., Zhang, F., Ban, R., Deussen,

O., and Wang, Y. (2023). Force-directed graph lay-

outs revisited: a new force based on the t-distribution.

IEEE Transactions on Visualization and Computer

Graphics.

The Interactive Network Visualization of the Interactions Between Topologically Associating Domains in the Genome of Fruit Fly

511